Databricks Machine Learning Associate Question 52
Single answerYou have trained multiple models using Databricks AutoML, and you would like to review the source code for the best-performing model. How can you locate the source code for the best model?
- A
Navigate to the AutoML experiment run in the MLflow UI, and download the 'notebook' artifact associated with the best model.
- B
Locate the best model by its accuracy metric in the MLflow UI, and click the 'Source Code' button to open the notebook directly.
- C
Open the Databricks AutoML dashboard and select the 'Source Code' tab for the best-performing model.
- D
In the AutoML-generated notebook, find the cell that references the best model and extract the source code from there.
Show answer and explanation
Correct answer: A
Explanation
Databricks AutoML automatically tracks experiment runs in MLflow, including the source code for the best-performing model as an artifact. To view the source code, navigate to the MLflow experiment run for the best model and download the 'notebook' artifact. This ensures transparency and reproducibility of the model development process.
- A. Correct.
Correct: The source code for the best-performing model can be found as a 'notebook' artifact in the associated MLflow experiment run. This is the correct method to access the source code.
- B. Incorrect.
Incorrect: While you can locate the best model in the MLflow UI, there is no 'Source Code' button to directly open the notebook.
- C. Incorrect.
Incorrect: The Databricks AutoML dashboard does not have a specific 'Source Code' tab. Source code is stored in the MLflow experiment as an artifact.
- D. Incorrect.
Incorrect: The AutoML-generated notebook provides details about the experiment, but it does not directly contain the source code for the best model.